The "gage" feature in Route_link.nc is the FID order, not the station id of "nudgingParams.nc"

[YM]

Dear friends,

I am trying to use the observed streamflow for nudging Data Assimilation in spin-up the WRF-Hydro model. The nudgingTimeSliceObs and nudgingParams.nc are well created by rwrfhydro function. The fixed-width 15 character string is also examined. The Rout_link.nc is created by the WRF-Hydro GIS Preprocessing Tools with the forcast *.csv file. The *.csv file is below:

FID	LON	LAT	Station_ID
1	-122.128	37.35812	5125
2	-122.135	37.42163	5101
3	-122.189	37.42338	11164500

I found the "gages" variable in the Rout_link.nc, is recorded with the FID list, 

"        "

"       1"

"       2"

"       3"

and it is not the same as Station ID in the nudgingParams.nc. like 

"                  5125"

"                  5101"

"           11164500"

I am not sure whether is the problem. Should I manually revise the Rout_link.nc file? Look forward to your answers. Thanks.

 

[Kevin Sampson]

YM,

This is not a problem. The Station_ID is not a column heading which is read by the GIS Pre-processing tools. You will also see the FID values (1, 2, 3) in the 'frxst_pts' gridded variable in the Fulldom_hires.nc file. Just keep track of the mapping between the FID values (which must be numbered 1...n) and your Station_ID values.

Thanks,

Kevin

Kevin Sampson, GISP
Associate Scientist
National Center for Atmospheric Research
P.O. Box 3000 Boulder, CO  80307-3000
Tel: 303.497.8121

--
You received this message because you are subscribed to the Google Groups "wrf-hydro_users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wrf-hydro_use...@ucar.edu.
To view this discussion on the web visit https://groups.google.com/a/ucar.edu/d/msgid/wrf-hydro_users/38707985-8b6b-4526-8a67-4a643809fb65%40ucar.edu.

[YM]

Dear Kevin,

Many thanks for the quick reply.

I am continuing to run the model as suggested. The expected results are computed but I noticed that there is a warning in the processing, e.g., "Ndg: WARNING Gages are apparently missing from the nudgingParams.nc file". I don't find problems on the nudgingParams.nc file. The nudgingParams.nc file is also attached for double check.

The nudgingDA might be not working as expected. I am not sure about this. Do you have any comments? I am also wondering to know why there is a warning for nudgingParams.nc. THANKS!

Thanks,

YM

* nudgingParams.nc

netcdf nudgingParams {
dimensions:
    stationIdStrLen = 15 ;
    stationIdInd = UNLIMITED ; // (3 currently)
    threshInd = 1 ;
    monthInd = 12 ;
    threshCatInd = 2 ;
variables:
    char stationId(stationIdInd, stationIdStrLen) ;
        stationId:units = "-" ;
        stationId:long_name = "USGS station identifer" ;
    float R(stationIdInd) ;
        R:units = "meters" ;
        R:long_name = "Radius of influence in meters" ;
    float G(stationIdInd) ;
        G:units = "-" ;
        G:long_name = "Amplitude of nudging" ;
    float tau(stationIdInd) ;
        tau:units = "minutes" ;
        tau:long_name = "Time tapering parameter half window size in minutes" ;
    float qThresh(stationIdInd, monthInd, threshInd) ;
        qThresh:units = "m^3/s" ;
        qThresh:long_name = "Discharge threshold category" ;
    float expCoeff(stationIdInd, monthInd, threshCatInd) ;
        expCoeff:units = "minutes" ;
        expCoeff:long_name = "Coefficient b in denominator e^(-dt/b)" ;
data:

 stationId =
  "             5101",
  "             5125",
  "         11164500" ;

 R = 0.25, 0.25, 0.25 ;

 G = 1, 1, 1 ;

 tau = 15, 15, 15 ;

 qThresh =
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100,
  -100 ;

 expCoeff =
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120,
  120, 120 ;
}

On Friday, August 2, 2019 at 4:13:18 PM UTC-6

, Kevin wrote:

To unsubscribe from this group and stop receiving emails from it, send an email to wrf-hyd...@ucar.edu.

[Molly]

Hi YM, 

Currently nudging is only available/supported in the national water model configuration of WRF-Hydro, though I see that your experimentation might be able to utilize it. Our Nudging method is documented in section 4 of the Technical Description. Please have a look at that.
Our subject matter expert in regards to this is out of the office, but might be able to take a look at your inquiry when he returns next week. 

Best regards,

Molly

[YM]

Dear Molly,

Thanks. YES. The nudging DA is only available in the NWM configuration of WRF-Hydro. The technical description is very helpful for my study. The Route_link.nc, nudgingParams.nc and nudgingTimeSliceObs are prepared as suggested. However, My primary concern is that the inconsistent implementation between "gages" feature in Route_link.nc and nudgingParams.nc/ nudgingTimeSliceObs blocks the running processes.

The warning in the model is "Ndg: WARNING Gages are apparently missing from the nudgingParams.nc file".

Your colleague Kevin (I guess) kindly reply that The "gages" feature is not a problem if traced from "frxst_pts" variable in the fulldom_hires.nc. I checked this suggestion again. It's ok.  but the warning remains displaying. That is my confusion. I don't know how to deal with it.

Sincerely,

YM

[jamesmcc]

Hello YM, 

Yes, you were right: the "gages" field in Route_link.nc needs to match the "stationId" in the nudingParam.nc file, including all whitespace (15char width, generally I use leading whitespace). All blank means "no gage".  The error you are getting means that the gages in your domain (Routelink) do not have parameters in the nudging parameters file. 

Ndg: WARNING Gages are apparently missing from the nudgingParams.nc file

Ndg: WARNING nGagesWParamsDom: 0

Ndg: WARNING nGagesDomain: 3

This is because, as you guessed, they gages in the routelink do not match the stationIds in nudgingParams. You can simply add the gages variable to the Route_link.nc, keeping track of the correct location (link or feature_id) to do so, that is the geolocation process of gages.

I personally never use frxst_pts output. I'm not going to comment on it, I dont think it's even available for NWM. But the NWM "CHANOBS" option will put out a similar file, but in netcdf format, that outputs only on (non-blank) gages in the routelink file. I'm not sure exactly how this is related.

Please also make sure that you create timeslice files for the assimilation or the nudging wont do anything. You can verify that nudging is working by plotting the values in the timeslices files against the model output. Note that the model time may not exactly match the observation time, so the modeled and observed values may not exactly match. But if the obs time is at the model time, then they should be within floating point precision.

I hope that helps,

James

[YM]

Dear James,

Your kindly comments are highly appreciated! I fully agree with your suggestions. YES. I should manually revise the "gages" feature so as to match the nudgingParams.nc file. I will try it. Should there be any concerns, will let you know. As for the comments of "CHANOBS" files, I know it and also use the information to examine the runoff simulation. It is beneficial. Thanks.

YM